Dexter Filmore wrote:
> Well, I've been telling for years now that we were better off with one desktop 
> that has the flexibility to adapt to everyone's needs.


I can't imagine when it was that "we were better off with one desktop". Was it 
perhaps in the days of the Windows 95 desktop? Or the Apple Mac System 6 Finder?


> Be lightweight without graphical mumbo jumbo if desired, be all the visual 
> monster with tons of effects, be as simple as a task bar and systray, be a 
> full blown cornucopia of gadgets if somebody prefer that. 

So what you are actually saying is that this one desktop should actually be... 
a hundred different desktops, only all in one code base.

I'm guessing that you're not a programmer or an engineer, are you?

Configuration choices increase the complexity of a program exponentially. One 
desktop capable of being all things to all people, as you suggest, would be a 
thousand times more complicated than 100 desktops which each focus on one 
small segment of the users. That means a thousand times bigger, a thousand 
times more development time, and a thousand times more bugs.

Desktops of the complexity of KDE or Gnome are, in my opinion, already at the 
edge of being too complex to be maintained successfully. With so many 
combinations of configuration options, there is no possible way that every 
combination has been fully tested and is bug free. The best we can hope for is 
that the most commonly used combinations are bug free, and that any bugs are 
buried in combinations that nobody uses.


> Make it configurable from simply to rocket science, from 486 to i7 but have 
> ONE API. Offer developers a safe base.

While you're wishing, don't forget to ask for a flying pony that craps rainbows.


[...]
> Desktop Environment developers reinvent the wheel over and over again.
> My favorite picture viewer is GThumb. It's GTK so it looks a wee bit different 
> from qt/kde no matter how much I adapt themes and engines.
> It's gui bahviour is gtk and I can't do much about it.
> I do *not* have a choice if I want to stick with that program. (Unless I port 
> it to qt myself. Some choice.)

But it *is* a choice, and the only reasonable choice. Why should the GThumb 
programmer spend hundreds of hours, perhaps thousands of hours, trying to 
adapt his program to every imaginable toolkit?

If *you* want to turn GThumb into a qt app, then *you* can do the work, or pay 
somebody to do it for you.


[...]
> What we don't need is NOT a more powerful desktop, what we need are more 
> powerful *programs*.

The desktop is a program. Many programs. Are you saying that they *don't* need 
to be more powerful? Who decides which programs are allowed to be more 
powerful and which are not?

I think the KDE 4 developers made an incredible boneheaded mistake in the way 
they abandoned KDE 3 and started a new project from scratch. I think that the 
new functionality they created is mostly unnecessary and mostly unusable. (If 
I thought the opposite, I would be using KDE 4.) But it was their right to 
make that mistake, and who knows, next time they might actually get it right.


[...]
> What we need is not a better desktop, that's like saying we need a better 
> hammer. 
> There is nothing to improve about hammers. The one I get in a hardware shop 
> has been perfected to its purpose.

And which hammer would that be?

Tack hammer.
Ball-peen hammer.
Cross-peen hammer.
Sledgehammer.
Drilling hammer.
Bush hammer.
Claw hammer.
Framing hammer.
Geologist's hammer.
Lump or mash hammer.
Rubber mallet.
Copper or lead mallet.
Wooden mallet.
Dead blow hammer.
Soft-faced hammer.
Stonemason's hammer.
Tinner's hammer.
Dog-head hammer.


It never ceases to amuse me when people use the hammer as an analogy for why 
we only need one tool for some purpose.


> Leave the desktops alone. All we need to organize our programs is a taskbar, 
> launchers and a systray. If at all.

Maybe that's all *you* need.




-- 
Steven